Skip to yearly menu bar Skip to main content


Workshop

DLVM: A modern compiler infrastructure for deep learning systems

Richard Wei · Lane Schwartz · Vikram Adve

East Meeting Level 8 + 15 #9

Thu 3 May, 11 a.m. PDT

Deep learning software demands reliability and performance. However, many of the existing deep learning frameworks are software libraries that act as an unsafe DSL in Python and a computation graph interpreter. We present DLVM, a design and implementation of a compiler infrastructure with a linear algebra intermediate representation, algorithmic differentiation by adjoint code generation, domain- specific optimizations and a code generator targeting GPU via LLVM. Designed as a modern compiler infrastructure inspired by LLVM, DLVM is more modular and more generic than existing deep learning compiler frameworks, and supports tensor DSLs with high expressivity. With our prototypical staged DSL embedded in Swift, we argue that the DLVM system enables a form of modular, safe and performant frameworks for deep learning.

Live content is unavailable. Log in and register to view live content